Intelligence and Miscellaneous Articles. 413 



In all these last experiments the magnetization was effected by 

 putting one extremity of the bar in contact with the pole of a 

 magnet ; but I have ascertained that the results are absolutely the 

 same when one extremity of the bar is introduced into an induction- 

 coil traversed by a current. 



These new facts cannot be explained in so simple a manner as 

 those which I previously made known (nos. 1, 2, 3). These latter 

 can be accounted for, as I have intimated, by considering only the 

 variations of the coercive force which result from the variations of 

 the temperature. When this is raised, the coercive force is lessened 

 and the magnetism will be augmented, supposing the magnetizing 

 force and the antagonistic molecular force to preserve sensibly the 

 same values between the limits of temperature considered. When 

 the bar is cooled the coercive force increases ; but as it is a passive 

 force, it cannot impress upon the molecules a movement inverse to 

 that which they executed during the heating. The acquired mag- 

 netism should therefore persist ; and we have seen (no. 2) that in 

 fact there is no retrogression during the cooling. This non-retro- 

 gression appears to me to characterize the coercive force ; for if that 

 force did not exist and the molecules were solicited exclusively by 

 the active forces, the magnetism would be the same during the 

 heating and during the cooling when the bar passed through the 

 same temperatures. 



To account for the facts nos. 4, 5, and 6, it appears to me indis- 

 pensable to have recourse to other considerations. ISTo. 4 would be 

 explained by admitting that the magnetizing force diminishes when 

 the temperature rises ; and this is a point generally admitted. But 

 to account for nos. 5 and 6 it seems necessary to introduce a new 

 hypothesis, which consists in supposing that, when the superficial 

 layers of a steel bar are heated sufficiently to be incapable of 

 receiving any further notable magnetization, they permit the deeper 

 layers (which were not magnetized at all at the ordinary tempera- 

 ture) to receive a certain degree of magnetization. This hypothesis 

 admitted, facts 5 and 6 are explained : the total magnetization of 

 the bar, after complete cooling, is the greater the more the bar has 

 been heated, because the number of layers which partake of the 

 magnetization increases with the temperature ; the superficial layers, 

 according to the hypothesis, are not sensibly magnetized as long as 

 the bar is strongly heated ; but they receive magnetism when the 

 bar is sufficiently cooled ; and this magnetism is added to that of 

 the deeper layers. 



If one admits that the magnetizing force exerted by the current 

 of an induction-coil upon a steel bar diminishes when the tempera- 

 ture of the bar rises above a certain limit, one can hardly refuse to 

 admit that reciprocally the inducing action exerted by the bar upon 

 the closed circuit of a coil surrounding it must equally diminish 

 when the temperature increases beyond the same limit. Therefore, 

 even if in this case the orientation of the molecules should be in- 

 variable, the demagnetizing current must diminish; and as this 

 current serves me for measuring the magnetization, the magnetiza- 

 tion must itself undergo an apparent diminution. This variation 



